RESUMO
OBJECTIVE: To develop an index to determine which opioid-exposed neonates have the most severe neonatal abstinence syndrome (NAS). STUDY DESIGN: Full-term neonates with NAS (n = 116) from mothers maintained on methadone or buprenorphine were enrolled from 8 sites into a randomized clinical trial of morphine vs methadone. Ninety-nine (85%) were evaluated at hospital discharge using the NICU Network Neurobehavioral Scale (NNNS). At 18 months, 83 of 99 (83.8%) were evaluated with the Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-III), and 77 of 99 (77.7%) were evaluated with the Child Behavior Checklist (CBCL). RESULTS: Cluster analysis was used to define high (n = 21) and low (n = 77) NAS severity. Compared with infants in the low NAS severity cluster, infants in the high NAS severity cluster had a longer length of stay (P < .001), longer length of stay due to NAS (P < .001), longer duration of treatment due to NAS (P < .001), and higher total dose of the study drug (P < .001) and were more likely to have received phenobarbital (P < .001), to have been treated with morphine (P = .020), and to have an atypical NNNS profile (P = .005). The 2 groups did not differ in terms of maximum Finnegan score. At 18 months, in unadjusted analyses, compared with the high-severity cluster, the low-severity cluster had higher scores on the Bayley-III Cognitive (P = .013), Language (P < .001), and Motor (P = .041) composites and less total behavior problems on the CBCL (P = .028). In adjusted analyses, the difference in the Bayley-III Language composite remained (P = .013). CONCLUSIONS: Presumptive measures of NAS severity can be aggregated to develop an index that predicts developmental outcomes at age 18 months.